摘要
本研究旨在利用近红外光谱(NIRS)技术构建芦竹全株中水分(MC)、粗蛋白质(CP)、粗脂肪(EE)、粗灰分(Ash)、粗纤维(CF)、中性洗涤纤维(NDF)、酸性洗涤纤维(ADF)和酸性洗涤木质素(ADL)含量的预测模型。本试验所用芦竹样品采自原阳和驻马店种植区,刈割高度分别为1.0、1.5、2.0和2.5 m。芦竹样品分为定标集(n=95)和验证集(n=31),通过NIRS技术结合主成分分析及改良偏最小二乘法(MPLS)等方法,建立样品特征光谱与待测化学成分之间的预测模型。结果表明:芦竹CP含量的外部验证相对分析误差(RPD)和预测决定系数(RSQ)最高,分别为10.82和0.991;EE、Ash、CF、NDF、ADF和ADL含量的RPD在4.0~6.3,其RSQ均大于0.94;MC含量的RPD和RSQ最低,分别为2.78和0.868。本试验所建芦竹MC、CP、EE、Ash、CF、NDF、ADF和ADL含量的预测模型有较好的预测能力,可应用于实际生产。
In this study,near infrared spectroscopy(NIRS)technology was used to construct a prediction model of water(MC),crude protein(CP),crude fat(EE),crude ash(Ash),crude fiber(CF),neutral detergent fiber(NDF),acid detergent fiber(ADF)and acid detergent lignin(ADL)contents in whole plant of Arunax arunax.The samples were taken from the planting areas of Yuanyang and Zhumadian,and the cutting heights were 1.0,1.5,2.0 and 2.5 m,respectively.The samples were divided into calibration set(n=95)and validation set(n=31).The prediction model between characteristic spectrum and chemical composition was established by NIRS technology combined with principal component analysis and modified partial least square method(MPLS).The results showed that the ratio of performance to deviation for validation(RPD)and the coefficient of determination for validation(RSQ)of CP content for Arundo donax were the highest,which were 10.82 and 0.991.The RPD of EE,Ash,CF,NDF,ADF and ADL were between 4.0 to 6.3,and their RSQ were all above 0.94.The RPD and RSQ of MC content were the lowest,which were 2.78 and 0.868.In conclusion,the prediction model establish in the present study can be used to predict the contents of MC,CP,EE,Ash,CF,NDF,ADF and ADL of Arundo donax,which can be applied to the actual production.
作者
王改利
王琳燚
陈京创
戴玉瑞
周莹
席燕燕
李绍钰
徐彬
WANG Gaili;WANG Linyi;CHEN Jingchuang;DAI Yurui;ZHOU Ying;XI Yanyan;LI Shaoyu;XU Bin(Animal Husbandry Institute,Henan Academy of Agricultural Sciences,Zhengzhou 450002,China;Henan Dayong Industrial Co.,Ltd.,Hebi 456750,China)
出处
《动物营养学报》
北大核心
2025年第3期2058-2065,共8页
CHINESE JOURNAL OF ANIMAL NUTRITION
基金
河南省农科院自主创新项目(2024ZC092)
河南省农业科学院科技创新团队专项(2024TD13)。
关键词
芦竹
近红外光谱技术
营养组分
预测模型
Arundo donax
near infrared reflectance spectroscopy technology
nutrient component
prediction model
